Vol. 6, 1920 
PHYSIOLOGY: ARM SB Y, FRIES AND BRA MAN 
265 
A computation of the closeness of fit of the Hnes represented by the fore- 
going equations to the individual results gives the following values for the 
index (A) and the probability (P). 
It would perhaps have been preferable to make the comparisons w^th the 
dry matter of the feed consumed. No succulent feeds, however, were 
contained in the rations and it does not seem probable that the slight 
differences in moisture content can have materially affected the conclu- 
sions. We have not attempted to distinguish between the effects of differ- 
ent kinds or classes of feeding stuffs and doubt whether our data are suffi- 
ciently extensive to warrant it. 
It would appear from the foregoing data that within the range of these 
experiments, i. e., with feed varying from 5 to 27 grams per kilogram live 
weight, equation (3) may be used for computing the CO2 : heat ratio of 
cattle when their live weight and the amount of feed consumed are known, 
and that by the use of this ratio the heat production may be computed 
from the observed CO2 production with a good degree of accuracy, at 
least for animals on dry feed and not performing work. 
It should be noted, however, that the values of computed from equa- 
tion (3) differ slightly from the ratios computed from the corresponding 
values of yi and y2, indicating that ys is not a strictly linear function of x. 
The discrepancy does not appear sufficient to affect materially the compu- 
tation of the heat production from the CO2 on the amounts of feed used 
in these experiments but for the fasting animal, i. e., when x = 0, the dis- 
crepancy is much more considerable, viz. : 
It appears probable that the values of ys would be represented more ex- 
actly by a line slightly convex toward the horizontal axis and curving some- 
what more sharply upward as the value of x approaches zero, and that for 
very small amounts of feed the CO2 : heat ratio computed from the 
values of yi and y2 is more accurate than that computed from equation 
(3). This conclusion is strengthened by the fact that the fasting value 
thus obtained approximates quite closely that observed by Benedict in 
his experiments on fasting men. 
^ The entire credit for the mathematical discussion of the results as well as for the 
laborious computations involved is due to Mr. Braman. 
2 In most instances two of these periods constituted the two halves of one of the 48- 
hour experiments mentioned subsequently. 
Ai = 4.489 
A2 = 2.467. 
A3 = 0.663 
Pi = 
P2 = 
Ps = 
0.9508 
0.9640 
0.9986 
= 3.25 
J2 
ys = 2.80 
